A portable acoustic imager field fast calibration method

Abstract

The application discloses a kind of portable acoustic imaging instrument field fast calibration method, belong to acoustic imaging technical field.The application is aimed at the problem that the precision of acoustic imaging instrument is deteriorated due to the drastic change of temperature and humidity under complex outdoor working conditions, and proposes the following steps: real-time acquisition of field temperature and humidity and background noise;The environment is divided into a preset temperature and humidity interval category and the matching fast calibration strategy is called;Emit the coded excitation sound wave suitable for the interval category and synchronously emit the optical reference signal;Synchronously collect acoustic and image data, and suppress interference using spatial filtering;According to the compensation model corresponding to the interval, correct the sound velocity, equalize in frequency domain and correct the channel consistency, and solve the calibration parameter set;Apply the parameters to beam forming and acoustic-optical fusion algorithm, and determine the calibration effectiveness through residual verification.The application realizes automatic fast calibration under field temperature and humidity mutation, effectively guarantees the accuracy and data reliability of sound source positioning and sound pressure level measurement in complex environment.

Description

Technical Field

[0001] This invention belongs to the field of acoustic imaging technology, specifically relating to a method for rapid on-site calibration of a portable acoustic imaging device. Background Technology

[0002] Portable acoustic imagers, as sound source visualization and detection devices based on microphone arrays and beamforming algorithms, have been widely used in industrial operation and maintenance scenarios such as power line inspection, industrial leak detection, and equipment noise diagnosis. Their working principle relies on high-precision acquisition of the time difference or phase difference of sound waves arriving at each element of the array, combined with known sound velocity parameters, to invert the spatial coordinates and sound pressure level distribution of the sound source using beamforming algorithms. To ensure measurement accuracy, existing technologies typically perform full-scale metrological calibration in a constant temperature and humidity laboratory environment using a standard sound source and a precision displacement platform, or perform single-point sensitivity verification and acoustic-optical alignment operations on-site using a portable acoustic calibrator. Such calibration procedures can meet routine testing needs in environments with stable conditions and controllable background noise, and are gradually being integrated into the equipment's standard operating procedures.

[0003] However, in complex outdoor conditions (such as extreme cold, high temperature, high humidity, or sudden diurnal temperature variations), the actual temperature and humidity often deviate significantly from the equipment's standard calibration conditions. Drastic changes in temperature and humidity directly alter the airborne sound velocity propagation reference, the high-frequency sound absorption coefficient, and the microphone diaphragm tension and array structure dimensions. This leads to phase consistency drift between array channels, sound velocity calculation model mismatch, and beamforming algorithm weight failure, resulting in increased sound source localization deviation and sound pressure level measurement distortion, among other accuracy degradation issues. Existing rapid on-site calibration methods are mostly designed based on constant environmental assumptions, lacking real-time sensing and dynamic compensation mechanisms for temperature and humidity parameters. Furthermore, the calibration process relies on fixed reference models or manual intervention, making it difficult to quickly reconstruct reliable acoustic calibration benchmarks in scenarios with sudden temperature and humidity changes. Therefore, there is an urgent need for a portable acoustic imager calibration method that can adapt to drastic outdoor temperature and humidity changes and automatically complete environmental parameter coupling compensation and rapid calibration before use, to ensure detection accuracy and data reliability under complex on-site conditions. Summary of the Invention

[0004] To address the aforementioned problems in the prior art, this invention provides a method for rapid on-site calibration of a portable acoustic imager.

[0005] The objective of this invention can be achieved through the following technical solutions: A method for rapid on-site calibration of a portable acoustic imager includes the following steps: S1: The environmental sensing module collects temperature and humidity parameters and background noise spectrum characteristics of the calibration site in real time. S2: Based on the temperature and humidity parameters, classify the current environment into a preset temperature and humidity range category, and call the fast calibration strategy that matches the range category; S3: Control the calibrator to emit coded excitation acoustic waves that are appropriate for the current interval category, and simultaneously emit optical positioning reference signals; S4: Acoustic response data and image reference data are acquired synchronously through the microphone array of the acoustic imager and the optical camera, and spatial filtering algorithm is used to suppress incoherent environmental interference. S5: Based on the compensation model corresponding to the temperature and humidity range category, perform sound velocity reference correction, frequency domain attenuation equalization and array channel consistency correction on the collected acoustic response data, and calculate the calibration parameter set. S6: Apply the calibration parameter set to the beamforming and acousto-optic fusion algorithm of the acoustic imager, and determine the validity of the calibration result through the residual verification mechanism. If valid, output the calibration report and enter the detection mode. As a further aspect of the present invention, the preset temperature and humidity range categories in S2 include a standard calibration range, a mild offset range, and a coupled extreme range; the division logic is as follows: when the temperature T∈[15℃,25℃] and the relative humidity RH∈[40%,60%], it is determined to be a standard calibration range; when T∈[5℃,15℃)∪(25℃,35℃] and RH∈[30%,70%], it is determined to be a mild offset range; when T>35℃ or T<5℃, and RH>70% or RH<30%, it is determined to be a coupled extreme range; for different range categories, an independent sound velocity model c(T,RH) and an air absorption coefficient model α(f,T,RH) are loaded respectively.

[0006] As a further aspect of the present invention, when the standard calibration zone or the mild offset zone is determined, the fast calibration strategy in S5 adopts a linear single-parameter compensation model: based on the real-time temperature T, the sound velocity reference value in the beamforming algorithm is dynamically updated according to the formula c=331.4+0.606·T, and the gain of each channel of the microphone array is linearly corrected according to the preset temperature sensitivity drift coefficient, and the first calibration parameter set containing the sound velocity correction factor and the channel equalization coefficient is calculated.

[0007] As a further aspect of the present invention, when the high-temperature drying coupling zone is determined, the rapid calibration strategy in S5 adopts a high-frequency attenuation compensation and electrostatic noise suppression model: the nitrogen molecule relaxation absorption peak frequency band is calculated based on real-time T and RH, and a pre-emphasis gain is applied to this frequency band in the frequency domain weight matrix of the encoded excitation sound wave; at the same time, by collecting the array background noise spectrum in the static state, the beamforming threshold value of this frequency band is dynamically raised, and the microphone phase lag caused by high-temperature drying is compensated, and a second calibration parameter set including frequency domain pre-equalization weight and phase advance compensation amount is calculated.

[0008] When the area is determined to be a high-temperature and high-humidity coupling zone, the rapid calibration strategy in S5 adopts a two-dimensional acousto-optic joint distortion compensation model: based on real-time T and RH, the two-dimensional sound velocity meter is called to compensate for the sound wave propagation delay. At the same time, the thermal and humid expansion deformation matrix of the array shell is inverted using the built-in inertial measurement unit and the laser ranging module to perform nonlinear geometric correction on the array element spatial coordinates; and the frequency band attenuation compensation is performed on the sound pressure level reading by combining the relaxation absorption characteristics of water molecules, and the third calibration parameter set containing the geometric deformation correction matrix and nonlinear acousto-optic mapping parameters is calculated.

[0009] As a further aspect of the present invention, when the low-temperature and high-humidity coupling zone is determined, the rapid calibration strategy in S5 adopts an anti-condensation dynamic calibration and oxygen relaxation compensation model: real-time monitoring of the equivalent impedance change rate of the microphone diaphragm, triggering the dynamic sensitivity recalibration process when the impedance change exceeds the threshold; for the mid-to-low frequency absorption enhancement caused by oxygen molecule relaxation, the steering vector weight of the corresponding frequency band is reduced in the beamforming algorithm; at the same time, the sound velocity reference is down-corrected based on the negative temperature coefficient, and a fourth calibration parameter set including the dynamic sensitivity correction coefficient and the mid-to-low frequency suppression factor is calculated.

[0010] As a further aspect of the present invention, the coded excitation sound waves in S3 and S4 are linear frequency modulated chirp signals or maximum length sequence (MLS) signals, with a sweep frequency range covering 5kHz-40kHz; the spatial filtering algorithm collects ambient noise on a reference microphone array arranged around the calibrator, estimates the main interference direction using the MUSIC algorithm, and constrains the interference direction gain to zero in the beamforming weight calculation to generate spatial nulls, thereby extracting coherent calibration signals in a non-anechoic environment.

[0011] As a further aspect of the present invention, the calibrator in S3 integrates a reference microphone with metrological traceability. When transmitting coded excitation sound waves, it synchronously acquires the actual output sound pressure level and dynamically adjusts the power amplifier gain through a closed-loop control circuit to stabilize the output sound pressure level within a preset standard value of ±0.5dB. The calibration report output in S6 includes the original waveform hash value of the reference microphone, environmental parameters, timestamp, and device serial number, forming an auditable digital traceability certificate.

[0012] As a further aspect of the present invention, the residual verification mechanism in S6 is specifically as follows: after the parameter application is completed, the calibrator is controlled to return to the reference position in front for retesting, and the pixel-level residual and sound pressure level deviation between the acoustic imaging center and the optical positioning reference are calculated; if the residual is greater than the preset threshold or the sound pressure level deviation exceeds ±1.5dB, it is determined that the calibration has failed, and the system is automatically switched to the compensation model of the adjacent temperature and humidity range or a second iteration calibration is triggered; if the calibration still fails after two consecutive iterations, the device is sent for inspection alarm.

[0013] As a further embodiment of the present invention, steps S3 to S5 are performed at the calibration site according to a five-point calibration array: coded excitation sound waves are repeatedly emitted and data is collected at positions directly in front of the acoustic imager, at horizontal ±15° and vertical ±10° positions; affine transformation and higher-order distortion correction parameters are calculated by solving the multi-point spatial coordinates, and the calibration parameter set is injected into the signal processing unit of the acoustic imager in the form of a configuration file to achieve real-time acousto-optic pixel-level alignment and positioning error compensation of subsequent detection data.

[0014] The beneficial effects of this invention are as follows: by subdividing the temperature and humidity environment into three major categories and multiple subcategories, the shortcomings of coarse-grained compensation are avoided. Each interval category is loaded with an independent sound velocity and air absorption model, making subsequent sound velocity benchmark correction and frequency domain attenuation equalization more accurate. In particular, for the coupled extreme regions, the interactive effects of temperature and humidity (such as the shift in nitrogen relaxation absorption peaks under high temperature and dryness, and the intensified relaxation of water molecules under high temperature and high humidity) are captured by independent models, significantly improving the adaptability of the calibration algorithm in harsh environments. Attached Figure Description

[0015] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0016] Figure 1 This is a flowchart of the steps of the present invention. Detailed Implementation

[0017] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided.

[0018] like Figure 1 As shown, a method for rapid on-site calibration of a portable acoustic imager includes the following steps: S1: The environmental sensing module collects temperature and humidity parameters and background noise spectrum characteristics of the calibration site in real time. S2: Based on the temperature and humidity parameters, classify the current environment into a preset temperature and humidity range category, and call the fast calibration strategy that matches the range category; S3: Control the calibrator to emit coded excitation acoustic waves that are appropriate for the current interval category, and simultaneously emit optical positioning reference signals; S4: Acoustic response data and image reference data are acquired synchronously through the microphone array of the acoustic imager and the optical camera, and spatial filtering algorithm is used to suppress incoherent environmental interference. S5: Based on the compensation model corresponding to the temperature and humidity range category, perform sound velocity reference correction, frequency domain attenuation equalization and array channel consistency correction on the collected acoustic response data, and calculate the calibration parameter set. S6: Apply the calibration parameter set to the beamforming and acousto-optic fusion algorithm of the acoustic imager, and determine the validity of the calibration result through the residual verification mechanism. If valid, output the calibration report and enter the detection mode. By acquiring real-time on-site temperature and humidity parameters and dynamically dividing the environment into intervals, this method overcomes the shortcomings of traditional methods that rely on the assumption of a constant environment. Different compensation models are applied to different intervals, avoiding error amplification caused by single parameter correction. Synchronous transmission of the coded excitation acoustic wave and optical reference signal ensures the consistency of the acoustic-optical joint calibration. The spatial filtering algorithm effectively suppresses background noise interference in non-anechoic environments, allowing the calibration process to be performed directly in complex outdoor conditions without returning to the laboratory. The residual verification mechanism ensures the reliability of the calibration results, preventing distortion of subsequent measurement data due to calibration failure. Overall, this method achieves rapid, automatic, and highly reliable calibration of portable acoustic imagers under conditions of drastic temperature and humidity changes.

[0019] As a further aspect of the present invention, the preset temperature and humidity range categories in S2 include a standard calibration range, a mild offset range, and a coupled extreme range; the division logic is as follows: when the temperature T∈[15℃,25℃] and the relative humidity RH∈[40%,60%], it is determined to be a standard calibration range; when T∈[5℃,15℃)∪(25℃,35℃] and RH∈[30%,70%], it is determined to be a mild offset range; when T>35℃ or T<5℃, and RH>70% or RH<30%, it is determined to be a coupled extreme range; for different range categories, an independent sound velocity model c(T,RH) and an air absorption coefficient model α(f,T,RH) are loaded respectively.

[0020] This embodiment, based on Embodiment 1, defines the preset temperature and humidity range categories in S2 as including a standard calibration zone, a mild offset zone, and a coupled extreme zone. The specific division logic is as follows: When the temperature T∈[15℃,25℃] and the relative humidity RH∈[40%,60%], it is determined to be the standard calibration zone, which is consistent with the factory calibration conditions of the equipment.

[0021] When T∈[5℃,15℃)∪(25℃,35℃] and RH∈[30%,70%], it is determined to be a mild offset zone. At this time, the temperature and humidity deviate from the calibration conditions but are still within the normal meteorological range.

[0022] When T>35℃ or T<5℃, and RH>70% or RH<30%, it is determined to be a coupling extreme zone, which is further subdivided into subcategories such as high temperature drying (T>35℃ and RH<30%), high temperature and high humidity (T>35℃ and RH>70%), and low temperature and high humidity (T<5℃ and RH>70%).

[0023] For different interval categories, independent sound speed models c(T,RH) and air absorption coefficient models α(f,T,RH) are loaded respectively. The sound speed model uses an empirical formula: c(T,RH)=331.4+0.606·T+0.012·(RH-50); In the formula, the second term is the temperature-dominant term, and the third term is the humidity correction term (the correction value is negative when RH < 50%). The baseline value for the speed of sound is 331.4 m / s (dry air at 0℃). The speed of sound increases by approximately 0.606 m / s for every 1℃ increase in temperature, and a small linear correction term is added when the humidity deviates from 50%. The air absorption coefficient model is based on the international standard ISO 9613-1, comprehensively considering classical absorption, rotational relaxation absorption, nitrogen relaxation absorption, and oxygen relaxation absorption. The functional relationships between each absorption component and frequency, temperature, and relative humidity are achieved through pre-stored lookup tables or polynomial fitting. A simplified model can be used for the standard calibration region, while the complete model needs to be called for the coupled extreme regions to account for nonlinear effects.

[0024] By subdividing the temperature and humidity environment into three main categories and multiple subcategories, the shortcomings of coarse-grained compensation are avoided. Each category is loaded with independent sound velocity and air absorption models, making subsequent sound velocity benchmark correction and frequency domain attenuation equalization more accurate. Particularly for the coupled extreme regions, the interactive effects of temperature and humidity (such as the shift in nitrogen relaxation absorption peaks under high temperature and dryness, and the intensified relaxation of water molecules under high temperature and high humidity) are captured by independent models, significantly improving the adaptability of the calibration algorithm to harsh environments.

[0025] As a further aspect of the present invention, when the standard calibration zone or the mild offset zone is determined, the fast calibration strategy in S5 adopts a linear single-parameter compensation model: based on the real-time temperature T, the sound velocity reference value in the beamforming algorithm is dynamically updated according to the formula c=331.4+0.606·T, and the gain of each channel of the microphone array is linearly corrected according to the preset temperature sensitivity drift coefficient, so as to obtain the first calibration parameter set containing the sound velocity correction factor and the channel equalization coefficient.

[0026] This formula is a classical empirical formula for the speed of sound, where 331.4 m / s is the speed of sound in dry air at 0℃, and 0.606 is the temperature coefficient (unit: m·s). - ¹·℃ - ¹), applicable to the 0~40℃ range. Simultaneously, the gain of each channel of the microphone array is linearly corrected according to a preset temperature sensitivity drift coefficient (e.g., -0.02 dB / ℃): Let the channel gain at the reference temperature T0=20℃ be G0, then the corrected gain at the current temperature is: G(T) = G0 + β·(T-T0); Where β is the drift coefficient, the first set of calibration parameters obtained by solving includes the sound speed correction factor (i.e., the ratio of c to the factory-calibrated sound speed c0) and the channel equalization coefficient (the gain adjustment amount of each channel).

[0027] This embodiment targets standard calibration areas with minimal environmental deviations and mild offset areas, employing a linear single-parameter compensation model. This model features low computational complexity and fast response, meeting the needs of rapid on-site calibration. The temperature-driven sound velocity correction formula is simple and reliable, requiring no humidity input and reducing sensor costs. Linear gain correction effectively compensates for sensitivity drift caused by microphone diaphragm tension changes with temperature, ensuring consistency between array channels. This method achieves lightweight calibration while maintaining accuracy, making it suitable for most typical outdoor scenarios.

[0028] As a further aspect of the present invention, when the high-temperature drying coupling zone is determined, the rapid calibration strategy in S5 adopts a high-frequency attenuation compensation and electrostatic noise suppression model: the nitrogen molecule relaxation absorption peak frequency band is calculated based on real-time T and RH, and a pre-emphasis gain is applied to this frequency band in the frequency domain weight matrix of the encoded excitation sound wave; at the same time, by collecting the array background noise spectrum in the static state, the beamforming threshold value of this frequency band is dynamically raised, and the microphone phase lag caused by high-temperature drying is compensated, and a second calibration parameter set including frequency domain pre-equalization weight and phase advance compensation amount is calculated.

[0029] This embodiment defines a rapid calibration strategy when the coupling zone is identified as a high-temperature, dry environment. In a high-temperature, dry environment (temperature above 35°C, humidity below 30%), the peak frequency of nitrogen molecule relaxation absorption shifts to the mid-to-high frequencies, with a typical center frequency between 50 and 100 kHz, resulting in significant attenuation of high-frequency energy within the acoustic imager's operating frequency band (5 to 40 kHz). The strategy employs a high-frequency attenuation compensation and electrostatic noise suppression model: First, the peak frequency of nitrogen molecule relaxation absorption is calculated based on real-time temperature and relative humidity. This peak frequency is calculated based on the relaxation time, which is inversely proportional to both temperature and humidity; the peak frequency equals the reciprocal of the relaxation time divided by twice pi. In the frequency domain weighting matrix of the encoded excitation sound wave, a pre-emphasis gain is applied to the frequency band near this peak frequency (e.g., the band within 0.7 to 1.3 times its value). The gain amount is determined inversely proportional to the difference between the absorption coefficient of this frequency band and that under ideal conditions. Simultaneously, by acquiring the array's background noise spectrum under static conditions, the beamforming threshold for that frequency band is dynamically increased (e.g., setting the threshold to three times the root mean square of the noise spectrum) to suppress electrostatic discharge noise that may be exacerbated by high-temperature and dry environments. Furthermore, compensation is provided for microphone phase lag caused by high-temperature and dry conditions: based on the microphone's equivalent circuit model, phase lag has an approximately linear relationship with temperature; therefore, a temperature-proportional lead compensation is set. The calculated second calibration parameter set includes the frequency domain pre-equalization weight matrix and the phase lead compensation.

[0030] To address the severe high-frequency attenuation and prominent electrostatic noise issues in high-temperature, dry coupling areas (such as deserts and outdoor substations in summer), frequency domain pre-emphasis and adaptive threshold suppression techniques were employed. By accurately calculating and compensating for the nitrogen relaxation absorption peak frequency band, the detectability of high-frequency sound sources was restored. Dynamic threshold suppression effectively improved the signal-to-noise ratio, preventing electrostatic noise from being misinterpreted as calibration signals. Phase lead compensation ensured the time delay accuracy of array beamforming. This strategy enables the acoustic imager to obtain accurate sound source localization and sound pressure level measurements even in high-temperature, dry environments.

[0031] As a further aspect of the present invention, when the high temperature and high humidity coupling zone is determined, the rapid calibration strategy in S5 adopts a two-dimensional acousto-optic joint distortion compensation model: based on real-time T and RH, the two-dimensional sound velocity meter is called to compensate for the sound wave propagation delay, and the thermal and humid expansion deformation matrix of the array shell is inverted using the built-in inertial measurement unit and the laser ranging module to perform nonlinear geometric correction on the array element spatial coordinates; and the frequency band attenuation compensation is performed on the sound pressure level reading by combining the relaxation absorption characteristics of water molecules, and the third calibration parameter set containing the geometric deformation correction matrix and nonlinear acousto-optic mapping parameters is calculated.

[0032] This embodiment defines a rapid calibration strategy when the environment is identified as a high-temperature and high-humidity coupling zone. In high-temperature and high-humidity environments (temperature above 35°C, humidity above 70%), air density decreases and water molecule content is extremely high, leading to complex changes in sound wave propagation delay. Simultaneously, the thermal and hygroscopic expansion of the array shell causes distortion of the array element spatial coordinates. The strategy employs a two-dimensional acousto-optic joint distortion compensation model: First, based on real-time temperature and relative humidity, a pre-generated two-dimensional sound velocity lookup table is used to compensate for the sound wave propagation delay. Second, the thermal and hygroscopic expansion deformation matrix of the array shell is inverted using the built-in inertial measurement unit and laser ranging module. Specifically, the array element coordinate matrix is ​​measured at the factory calibration temperature and humidity as a reference. Under current conditions, the shell exhibits linear thermal expansion and hygroscopic expansion coefficients, causing the array element coordinates to change proportionally. Due to the anisotropy of the shell material and uneven local humidity distribution, the displacement of multiple reference points needs to be measured using the laser ranging module to invert a nonlinear geometric correction matrix. Multiplying the current coordinates by this matrix yields the corrected coordinates. Then, the sound pressure level readings are compensated for by frequency band attenuation based on the relaxation absorption characteristics of water molecules (whose peak frequency is approximately 20 to 50 kHz). The compensation coefficient is exponentially related to the absolute humidity. The calculated third calibration parameter set includes the geometric deformation correction matrix and nonlinear acousto-optic mapping parameters, such as the offset of the optical camera's intrinsic parameters caused by shell deformation.

[0033] To address the two major challenges of array geometric distortion and acoustic propagation nonlinearity in high-temperature and high-humidity environments (such as tropical rainforests and coastal areas in summer), a two-dimensional acousto-optic joint compensation scheme is proposed. By using an inertial measurement unit and laser ranging to invert the deformation matrix in real time, the problem of the traditional rigid array assumption failing is solved. Water molecule relaxation compensation ensures the accuracy of sound pressure level measurement. This strategy, for the first time, decouples and jointly compensates for geometric deformation and acoustic propagation effects, significantly improving the positioning and quantitative accuracy in high-temperature and high-humidity environments.

[0034] As a further aspect of the present invention, when the low-temperature and high-humidity coupling zone is determined, the rapid calibration strategy in S5 adopts an anti-condensation dynamic calibration and oxygen relaxation compensation model: real-time monitoring of the equivalent impedance change rate of the microphone diaphragm, triggering the dynamic sensitivity recalibration process when the impedance change exceeds the threshold; for the mid-to-low frequency absorption enhancement caused by oxygen molecule relaxation, the steering vector weight of the corresponding frequency band is reduced in the beamforming algorithm; at the same time, the sound velocity reference is down-corrected based on the negative temperature coefficient, and a fourth calibration parameter set including the dynamic sensitivity correction coefficient and the mid-to-low frequency suppression factor is calculated.

[0035] This embodiment defines a rapid calibration strategy when the environment is identified as a low-temperature, high-humidity coupling zone. In low-temperature, high-humidity environments (temperature below 5°C, humidity above 70%), condensation easily occurs on the microphone diaphragm surface, leading to an increase in the diaphragm's equivalent mass and a decrease in compliance, manifesting as a sudden impedance change. The strategy employs an anti-condensation dynamic calibration and oxygen relaxation compensation model: real-time monitoring of the microphone diaphragm's equivalent impedance change rate. The impedance detection method involves applying a known small-amplitude high-frequency test signal to the diaphragm and detecting the response current. The impedance change rate is defined as the difference between the current impedance and the non-condensation reference impedance divided by the reference impedance. When the impedance change rate exceeds a preset threshold (e.g., 10%), a dynamic sensitivity recalibration process is triggered: the built-in reference sound source emits a standard amplitude (e.g., 94 dB @ 1 kHz) sound signal, and the sensitivity correction coefficient is calculated based on the current output voltage to update the sensitivity parameters of each channel. Meanwhile, to address the enhanced absorption in the mid-to-low frequencies caused by oxygen molecule relaxation (typical peak values ​​range from tens to hundreds of hertz, but their tailing effect can reach thousands of hertz), the steering vector weights for the corresponding frequency bands (e.g., 500 hertz to 4 kilohertz) are reduced in the beamforming algorithm to suppress spurious peaks caused by excessive absorption. The sound velocity reference is downgraded based on a negative temperature coefficient: using a temperature-dominated sound velocity formula (331.4 plus 0.606 multiplied by temperature), a sound velocity value below 331.4 m / s is naturally obtained when the temperature is below 5°C. The calculated fourth calibration parameter set includes dynamic sensitivity correction coefficients and mid-to-low frequency suppression factors.

[0036] To address the two prominent issues of condensation and oxygen relaxation absorption in low-temperature and high-humidity environments (such as winter smog and cold storage inspections), an active monitoring and dynamic recalibration mechanism is proposed. Condensation is detected in real time by impedance change rate, avoiding sensitivity drift caused by diaphragm quality changes. Oxygen relaxation compensation maintains the positioning accuracy of mid- and low-frequency sound sources. This strategy endows the acoustic imager with the ability to self-recover and recalibrate in humid and low-temperature environments, significantly improving the field usability of the equipment.

[0037] As a further aspect of the present invention, the coded excitation sound waves in S3 and S4 are linear frequency modulated chirp signals or maximum length sequence (MLS) signals, with a sweep frequency range covering 5kHz-40kHz; the spatial filtering algorithm collects ambient noise on a reference microphone array arranged around the calibrator, estimates the main interference direction using the MUSIC algorithm, and constrains the interference direction gain to zero in the beamforming weight calculation to generate spatial nulls, thereby extracting coherent calibration signals in a non-anechoic environment.

[0038] This embodiment defines the specific implementation of the coded excitation acoustic wave and spatial filtering algorithm in S3 and S4. The coded excitation acoustic wave is a linear frequency modulated (Chirp) signal or a maximum length sequence (MLS) signal. A linear frequency modulated signal is a signal whose frequency varies linearly with time, and its sweep range covers 5 kHz to 40 kHz. This signal has a finite duration in the time domain, a constant amplitude, and its instantaneous frequency increases linearly from the starting frequency to the ending frequency. A maximum length sequence signal is a periodic pseudo-random binary sequence whose autocorrelation function approximates the impulse function, making it very suitable for measuring the impulse response of a system. The spatial filtering algorithm collects ambient noise by placing a reference microphone array (e.g., four reference microphones distributed around the calibrator) around the calibrator and estimates the main interference direction using a multiple signal classification algorithm. This algorithm performs eigenvalue decomposition based on the covariance matrix of the array received data, dividing the space into a signal subspace and a noise subspace, and obtains the angle of arrival of the interference source through spectral peak search. Then, in the beamforming weight calculation, a linear constraint minimum variance criterion is adopted: two constraints are set—the array response in the true direction of the calibrator must be 1 (ensuring lossless passage of the desired signal), and the array response in the estimated interference direction must be 0 (completely suppressing the signal in that direction). Under these constraints, the weight vector that minimizes the output power is solved, thereby generating a spatial null. In this way, coherent calibration signals can be effectively extracted in a non-anechoic environment.

[0039] By selecting linear frequency modulated (LFM) signals or maximum length sequence (MLS) signals as the encoding excitation, excellent time-frequency characteristics were obtained. LFM signals offer advantages such as high pulse compression gain and strong multipath resistance, while MLS signals perform well at low signal-to-noise ratios. The multi-signal classification spatial filtering algorithm can dynamically identify the main direction of background noise sources and generate nulls in outdoor non-anechoic environments, effectively suppressing incoherent interference such as traffic noise, wind noise, and industrial noise. This results in purer calibration signals acquired by the microphone array and improves the signal-to-noise ratio of the calculated calibration parameters. This is crucial for rapid on-site calibration, as users cannot guarantee that the calibration environment meets laboratory-level quiet conditions.

[0040] As a further aspect of the present invention, the calibrator in S3 integrates a reference microphone with metrological traceability. When transmitting coded excitation sound waves, it synchronously acquires the actual output sound pressure level and dynamically adjusts the power amplifier gain through a closed-loop control circuit to stabilize the output sound pressure level within a preset standard value of ±0.5dB. The calibration report output in S6 includes the original waveform hash value of the reference microphone, environmental parameters, timestamp, and device serial number, forming an auditable digital traceability certificate.

[0041] This embodiment defines the self-calibration and data traceability mechanism of the calibrator in S3. The calibrator integrates a metrologically traceable reference microphone (e.g., a half-inch prepolarized condenser microphone with a calibration certificate traceable to national standards). When transmitting coded excitation sound waves, the reference microphone synchronously acquires the actual output sound pressure level and dynamically adjusts the power amplifier gain through a closed-loop control loop to stabilize the output sound pressure level within ±0.5 dB of a preset standard value. The closed-loop control uses a proportional-integral-derivative algorithm, and the error signal is the difference between the set sound pressure level and the measured value from the reference microphone. Simultaneously, the calibration report output in S6 includes the following auditable information: the hash value of the original waveform of the reference microphone (used to verify that the calibration waveform has not been tampered with; the hash algorithm uses the SHA-256 standard), environmental parameters (temperature, humidity, atmospheric pressure), timestamp, and device serial number, forming a digital traceability certificate.

[0042] By incorporating a built-in reference microphone and closed-loop gain control, the accuracy and stability of the calibrator's output sound pressure level are ensured, eliminating output errors caused by battery voltage fluctuations, temperature drift, or component aging. The calibration report includes traceability information such as hash values, enabling independent auditing of each on-site calibration and complying with quality management system requirements such as ISO / IEC 17025. In the event of measurement disputes, the compliance of the calibration process can be determined by verifying the hash values ​​and environmental parameters, enhancing the data credibility of the portable acoustic imager in serious scenarios such as forensic identification and accident investigation.

[0043] As a further aspect of the present invention, the residual verification mechanism in S6 is specifically as follows: after the parameter application is completed, the calibrator is controlled to return to the reference position in front for retesting, and the pixel-level residual and sound pressure level deviation between the acoustic imaging center and the optical positioning reference are calculated; if the residual is greater than the preset threshold or the sound pressure level deviation exceeds ±1.5dB, it is determined that the calibration has failed, and the system is automatically switched to the compensation model of the adjacent temperature and humidity range or a second iteration calibration is triggered; if the calibration still fails after two consecutive iterations, the device is sent for inspection alarm.

[0044] This embodiment defines the specific implementation of the residual verification mechanism in S6. After the parameter application is completed, the calibrator is controlled to return to the reference position directly in front (e.g., the center point 1 meter away from the imager) for retesting. Two residuals are calculated: the first is the pixel-level residual between the acoustic imaging center and the optical positioning reference, i.e., the Euclidean distance between the pixel coordinates of the sound source center located by the acoustic imaging algorithm and the pixel coordinates of the reference spot center captured by the optical camera; the second is the sound pressure level deviation, i.e., the absolute difference between the standard sound pressure level output by the calibrator and the measured value by the acoustic imager. If the pixel residual is greater than a preset threshold (e.g., 5 pixels) or the sound pressure level deviation exceeds ±1.5 dB, the calibration is determined to be a failure. After failure, the system automatically switches to the adjacent temperature and humidity range compensation model (e.g., if the current standard calibration area is exceeded but the residual exceeds the standard, then the mild offset area model is tried), or a second iteration calibration is triggered (steps S3 to S5 are re-executed). If the calibration still fails after two consecutive iterations, a device inspection alarm is output, prompting the user to return the device to the repair center for in-depth calibration or hardware troubleshooting.

[0045] The validity of calibration parameters is ensured through a retesting and verification mechanism. The residual threshold setting (approximately 0.5 degrees of angular resolution for 5 pixels at the pixel level, and ±1.5 dB for the sound pressure level, meeting the tolerance requirements of IEC 61672 for sound level meters) is neither too stringent, leading to frequent retries, nor too lenient, masking faults. Automatic switching between adjacent interval models or iterative calibration provides fault tolerance, adapting to instantaneous fluctuations in the field environment. A test alarm is output after consecutive failures, preventing the equipment from operating with defects and protecting user interests. This mechanism forms a closed loop of "calibration-verification-recalibration," significantly improving the robustness of field calibration.

[0046] As a further embodiment of the present invention, steps S3 to S5 are performed at the calibration site according to a five-point calibration array: coded excitation sound waves are repeatedly emitted and data is collected at positions directly in front of the acoustic imager, at horizontal ±15° and vertical ±10° positions; affine transformation and higher-order distortion correction parameters are calculated by solving the multi-point spatial coordinates, and the calibration parameter set is injected into the signal processing unit of the acoustic imager in the form of a configuration file to achieve real-time acousto-optic pixel-level alignment and positioning error compensation of subsequent detection data.

[0047] This embodiment defines the detailed process of S3 to S5 performed on-site using a five-point calibration array. The five calibration points are, in order: directly in front of the acoustic imager (azimuth 0 degrees, pitch 0 degrees), 15 degrees to the left horizontally, 15 degrees to the right horizontally, 10 degrees to the top vertically, and 10 degrees to the bottom vertically. At each position, the calibrator repeatedly transmits coded excitation sound waves (e.g., 3 times per position and averaged), and the acoustic imager simultaneously acquires data. Affine transformation and higher-order distortion correction parameters are calculated using multi-point spatial coordinates. Specifically, given the five coordinate points of the calibrator in the world coordinate system and the corresponding estimated coordinates obtained by the acoustic imager through sound source localization, an affine transformation matrix (including rotation, scaling, and translation parameters) is solved to make the estimated coordinates approximate the true coordinates after transformation. For nonlinear distortions (such as lens barrel distortion or array installation errors), a higher-order polynomial correction model is used: the power combination of pixel coordinates (including cross terms, square terms, etc.) is used as the independent variable, the world coordinates are used as the dependent variable, and the polynomial coefficients are fitted using the least squares method. Finally, the calibration parameter set (including sound velocity correction, channel equalization, geometric correction matrix, distortion correction coefficient, etc.) is injected into the signal processing unit of the acoustic imager in the form of a configuration file (such as JSON or XML format) to achieve real-time acousto-optic pixel-level alignment and positioning error compensation of subsequent detection data.

[0048] By using a five-point calibration array to collect data from multiple angles in space, more systematic error parameters can be calculated compared to single-point calibration. Affine transformation corrects for rotation, scaling, and shearing distortions between the array and the camera, while higher-order polynomials compensate for lens nonlinear distortion and array flatness errors. Calibration parameters are injected into the hardware processing unit as a configuration file, avoiding online calibration for each detection and ensuring real-time performance. This method enables portable acoustic imagers to achieve consistent acousto-optic alignment accuracy across the entire field of view after rapid on-site calibration, significantly improving the accuracy and stability of sound source localization in complex scenes.

[0049] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for rapid on-site calibration of a portable acoustic imager, characterized in that: It includes the following steps: S1: Collect the temperature and humidity parameters and the background noise spectrum characteristics of the calibration site in real time through the environmental perception module; S2: Based on the temperature and humidity parameters, divide the current environment into preset temperature and humidity interval categories, and call the fast calibration strategy matching the interval category; S3: Control the calibrator to emit coded excitation sound waves adapted to the current interval category, and synchronously emit an optical positioning reference signal; S4: Synchronously collect acoustic response data and image reference data through the microphone array of the acoustic imager and the optical camera, and use the spatial filtering algorithm to suppress incoherent environmental interference; S5: According to the compensation model corresponding to the temperature and humidity interval category, perform sound speed reference correction, frequency-domain attenuation equalization and array channel consistency correction on the collected acoustic response data, and solve to obtain a calibration parameter set; S6: Apply the calibration parameter set to the beamforming and acoustic-optic fusion algorithms of the acoustic imager, and determine the effectiveness of the calibration result through the residual verification mechanism. If it is effective, output a calibration report and enter the detection mode.

2. The method for rapid on-site calibration of a portable acoustic imager according to claim 1, characterized in that: The division logic of the preset temperature and humidity interval categories in S2, including the standard calibration area, the mild offset area and the coupled extreme area, is as follows: when the temperature T ∈ [15°C, 25°C] and the relative humidity RH ∈ [40%, 60%], it is determined as the standard calibration area; when 5 ≤ T < 15 or 25 < T ≤ 35 and RH ∈ [30%, 70%], it is determined as the mild offset area; when T > 35°C or T < 5°C, and RH > 70% or RH < 30%, it is determined as the coupled extreme area; independent sound speed models c(T, RH) and air absorption coefficient models α(f, T, RH) are loaded for different interval categories respectively.

3. The method for rapid on-site calibration of a portable acoustic imager according to claim 2, characterized in that: When it is determined as the standard calibration area or the mild offset area, the fast calibration strategy in S5 adopts a linear single-parameter compensation model: based on the real-time temperature T, dynamically update the sound speed reference value in the beamforming algorithm according to the formula c = 331.4 + 0.606·T, and linearly correct the gains of each channel of the microphone array according to the preset temperature sensitivity drift coefficient, and solve to obtain a first calibration parameter set including the sound speed correction factor and the channel equalization coefficient.

4. The method for rapid on-site calibration of a portable acoustic imager according to claim 3, characterized in that: When it is determined as the high-temperature and dry coupling area, the fast calibration strategy in S5 adopts a high-frequency attenuation compensation and electrostatic noise suppression model: calculate the peak frequency band of nitrogen molecule relaxation absorption according to the real-time T and RH, and apply pre-emphasis gain to this frequency band in the frequency-domain weight matrix of the coded excitation sound wave; at the same time, by collecting the array background noise spectrum in the static state, dynamically raise the beamforming threshold value of this frequency band, and compensate for the microphone phase lag caused by high temperature and dryness, and solve to obtain a second calibration parameter set including the frequency-domain pre-equalization weight and the phase lead compensation amount.

5. The method for rapid on-site calibration of a portable acoustic imager according to claim 4, characterized in that: When it is determined as the high-temperature and high-humidity coupling area, the fast calibration strategy in S5 adopts a two-dimensional acoustic-optic joint distortion compensation model: based on the real-time T and RH, call the two-dimensional sound speed table to compensate the sound wave propagation delay, and at the same time use the built-in inertial measurement unit and laser ranging module to invert the thermal and humidity expansion deformation matrix of the array housing, and perform non-linear geometric correction on the array element spatial coordinates; In addition, the sound pressure level readings are compensated for by frequency band attenuation by combining the relaxation absorption characteristics of water molecules, and a third calibration parameter set containing geometric deformation correction matrix and nonlinear acousto-optic mapping parameters is obtained.

6. The method for rapid on-site calibration of a portable acoustic imager according to claim 5, characterized in that: When the coupling zone is determined to be low temperature and high humidity, the rapid calibration strategy in S5 adopts an anti-condensation dynamic calibration and oxygen relaxation compensation model: real-time monitoring of the change rate of the equivalent impedance of the microphone diaphragm, and triggering the dynamic sensitivity recalibration process when the impedance change exceeds the threshold; in response to the enhanced absorption in the mid and low frequencies caused by oxygen molecule relaxation, the steering vector weight of the corresponding frequency band is reduced in the beamforming algorithm; at the same time, the sound velocity reference is down-corrected based on the negative temperature coefficient, and the fourth calibration parameter set containing the dynamic sensitivity correction coefficient and the mid and low frequency suppression factor is calculated.

7. The method for rapid on-site calibration of a portable acoustic imager according to claim 6, characterized in that: The coded excitation sound waves in S3 and S4 are linear frequency modulated chirp signals or maximum length sequence (MLS) signals, with a sweep frequency range covering 5kHz-40kHz. The spatial filtering algorithm collects ambient noise on a reference microphone array arranged around the calibrator, estimates the main interference direction using the MUSIC algorithm, and constrains the interference direction gain to zero in the beamforming weight calculation to generate spatial nulls, thereby extracting coherent calibration signals in a non-anechoic environment.

8. The method for rapid on-site calibration of a portable acoustic imager according to claim 7, characterized in that: The calibrator in S3 integrates a metrologically traceable reference microphone, which synchronously acquires the actual output sound pressure level when transmitting coded excitation sound waves. The power amplifier gain is dynamically adjusted through a closed-loop control circuit to stabilize the output sound pressure level within a preset standard value of ±0.5dB. The calibration report output in S6 includes the original waveform hash value of the reference microphone, environmental parameters, timestamp, and device serial number, forming an auditable digital traceability certificate.

9. A method for rapid on-site calibration of a portable acoustic imager according to claim 8, characterized in that: The residual verification mechanism in S6 is as follows: after the parameter application is completed, the calibrator is controlled to return to the reference position in front for retesting, and the pixel-level residual and sound pressure level deviation between the acoustic imaging center and the optical positioning reference are calculated; if the residual is greater than the preset threshold or the sound pressure level deviation exceeds ±1.5dB, it is determined that the calibration has failed, and the system automatically switches to the compensation model of the adjacent temperature and humidity range or triggers a second iteration calibration; if it still fails after two consecutive iterations, the device is sent for inspection alarm.

10. A method for rapid on-site calibration of a portable acoustic imager according to claim 9, characterized in that: S3 to S5 are performed at the calibration site according to a five-point calibration array: coded excitation sound waves are repeatedly emitted and data is collected at positions directly in front of the acoustic imager, at horizontal ±15° and vertical ±10° positions; affine transformation and higher-order distortion correction parameters are calculated by solving the multi-point spatial coordinates, and the calibration parameter set is injected into the signal processing unit of the acoustic imager in the form of a configuration file to achieve real-time acousto-optic pixel-level alignment and positioning error compensation of subsequent detection data.

Images